Overvotes in New York City: Machines Need Usability, Too


[Photo by Katie Orlinsky courtesy of the New York Times]

This post has been updated to correct the name of co-author Sundeep Iyer. My apologies, Sundeep – if I’m going to recommend that people remember your name it would help to get it right in the first place!

Earlier this week, the Brennan Center at NYU released a study of overvotes in the 2010 New York election. Design Deficiencies and Lost Votes – co-authored by Larry Norden (a name you likely know) and Sundeep Iyer (who you, like me, should make a point to remember) – found that tens of thousands of votes had been invalidated due to voters’ casting more than one vote in a given race.

These findings are especially important because:

  1. in New York, 2010 was the first election where voters had encountered new optical scan voting machines;
  2. these problems disproportionately affected communities with lower-income, immigrant and nonwhite voters; and
  3. the problem is likely to recur in larger numbers in 2012 when many more voters unfamiliar with the new machines cast ballots for the first time.

Readers of this blog who have seen my posts on usability know that we have already discussed the need for ballot language and online resources like websites to be readable and usable for voters. Some of those same problems were at issue here; the ballots in 2010 included contests that spanned two rows and a simple instruction to “vote for one” – and while the report doesn’t mention it, neither the state nor the city’s elections websites are very helpful about how to properly mark a ballot.

Yet in this case, the biggest culprits may have been the machines themselves – at least in how they were programmed (as required by federal law) to notify voters of overvotes and give them an opportunity to correct the error. In New York, however, Brennan’s report found that the message was confusing and gave voters mixed messages about how to proceed (Appendix A, p. 27). They propose a revision modeled on language in jurisdictions with lower overvotes. Fortunately, it appears that New York will be moving (albeit with encouragement from the courts) to correct these problems in time for the 2012 general election.

This is good news; after all, to paraphrase political scientist Mark Lindeman, “isn’t it bizarre that we certify the vote tabulators to (supposedly) one vote in 500,000 vote positions, yet tolerate ballot and [interface] designs whose results are not always accurate to one intended vote in three?” Here’s hoping New York follows through – and properly – on addressing the problems identified in Brennan’s report.

BONUS NOTE FOR ELECTION GEEKS: The maps and statistics in this report are terrific … in particular, the map on p. 11 which highlights one “hot spot” at a polling place in the South Bronx – is fascinating. Check it out – and kudos to Larry Norden and Sundeep Iyer for a report that’s absolutely jammed with data for the geek on your holiday shopping list.

Be the first to comment on "Overvotes in New York City: Machines Need Usability, Too"

Leave a comment

Your email address will not be published.